260
PAMELA LABiPLUOB ROBINSON : TEE iWESOZOfC BSSURES Op
"El
THE MESOZOIC FISSURES OF THE BRISTOL CHANNEL AREA AND
T€EI'R VERTEBRATE FAUNAS. By PAMELA
LAMPLUGH
ROBINSON,
B.Sc.,
Department of Zoology, University College, London.
Professor P. B. MEDAWAR,F.R.S., F.L.S.)
(Communicated by
(With Plates 1 to 1 and 6 text-figures.)
[Read 19 Januwy 1956.1
CONTENTS.
Introduction ........................................................
HistoryofResearch ................................................
Fissures ............................................................
The Geology of the Mendip Area ......................................
The Fissures ofthe Mendip Area ......................................
The Age of the Triassic Fissure Deposits. .............................
Late Triassic Climate and Vegetation in the Bristol Channel Area. . . . . . . . . . . .
The A4ccumulation of the Bones in the Upper Triassic Underground
.
.
................................................
The Upper Triaasie Vertebrate Fauns ..................................
Acknowledgments ..................................................
Summary ..........................................................
References .
..........
....................................
Appendix I. Fossils obtained from the Gurney Slsde fissure ..............
Appendix 11. Fossils obtained from the silicified rock marked as Liaa on the
Old Series one inch Geological Survey map no. 19 ....................
Explwation of Plates ................................................
Page
260
261
164
264
266
271
275
277
277
278
279
279
280
281
281
INTRODUCTION.
Mesozoic deposits occurring as fissure m n g s in the Carboniferous Limestone
of the Bristol Channel area (Somerset, South Gloucestershire, Glamorgan) have
provided a rich yield of a great variety of terrestrial vertebrates that fall into two
groups. One, which consists mainly of mammals and mammal-like reptiles, is
Rhmtic or Lower Liassic in age, and is found in fissures of submarine origin. The
other group contains Rauropsid reptiles only, occurring in a matrix which may be
red, green, or yellow in colour and is devoid of marine invertebrates. It has been
suggested that the latter group is Triassic in age, possibly Keuper.
Most of the Rhaetic and Liassic terrestrial vertebrates found in this area have
already been described (Owen, 1871 ; Simpson, 1928 ; Parrington, 1941, 1946 ;
Kiihne, 1949,1950). Kiihne has described two of the submarine fissures, one Rhaetic
and one Liassic, in which he found his fossil material. The Triassic reptiles await
description, and their study will necessarily take several years. The fissures in which
these latter reptiles are found have not been described ; they are quite different
in origin and history from those of Rhaetic and Liassic age. No evidence has ever
been given for t'he age assigned to the Triassic fissure deposits, and on theoretical
grounds they could be of any age post-Carboniferousbut pre-Rhaetic.
This paper is intended to serve as a background to the descriptions of the various
Triassic reptiles from fissure localities in the Bristol Channel area. After a brief
description of their mode of origin and infilling the age of the fissure deposits is
BRISTOL CHANNEL AREA AND THEW VERTEBUTB FAUNAS
261
discussed and the probable physical conditions which obtained a t the time outlined.
A preliminary discussion of the whole of this Triassic,fauna, and its significance,
follows.
HISTORY
OF RESEARCH.
Charles Moore was a very active worker in the Bristol Channel area during the
nineteenth century. He was the first to recognize and examine systematically
certain deposits of Rhaetic and Liassic age occurring as fillings in fissures in the
Carboniferous Limestone near Frome and Bristol (Moore, 1867, 1881). His best
known work wa.s the discovery of twenty-nine mammalian teeth in a Rhaetic fissure
near Holwell (Owen, 1871 ; Simpson, 1928).
Mr. F. G. Hudson was the h t to discover .$ fossiliferous fissure of non-marine
origin. It was found in the Carboniferous Limestone a t Slickstones Quarry, near
CromhalI, Gloucestershire, in 1938. A few pieces of the material were sent to the
British Mueeum. They were briefly described by Dr. W. E. Swinton (1939), who
gave the name ~levosuurushudsoni t o jaws which appeared to be rhynchocephalian.
The bulk of the material, still undescribed, is in the possession of Professor W. F.
Whittard at Bristol.
Dr. W. G. Kiihne came t o England from Germany in 1939, with a particular
interest in Mesozoic mammals. The known material of Jurassic mammals consisted
mainly of teeth and jaws. Kiihne sought new and more adequate material, in
order to gain a better knowledge of mammalian evolution during Jurassic times.
He knew that a part of the Mendip Hills had remained a land area, as a series of
islands, in Rhaetic and early Jurassic times, that Moore had found mammalian
teeth in a submarine Rhaetic fissure in the area, and that a number of other fissures
had been described by Moore. The fissures had d l occurred in Carboniferous Limestone, and exposures of the Limestone were numerous in quarries and as natural
outcrops. On the basis of this knowledge, and realizing that no one had studied the
Mendip fissures since Moore died in 1881, Kiihne commenced a systematic search for
fossiliferous &sure-deposits in the Mendip area.
At Holwell, in Moore's ' Microlestes ' quarry (fig. l), he found twenty mammalian
teeth in a fissure-mng (Kiihne, 1946 ; Parrington, 1941, 1946). Most of these teeth
were Microleptids, but two were Triconodonts, the oldest record of the group then
known. Near Shepton Mallet, a t Winhor Hill Quarry (fig. l), he discovered a
fissure of Lower L i m i c age contajnjng abundant bones of Oligokyphus, previously
known only from teeth, and h a been able to write a fairly complete osteology of the
whole animal. In 1946, at Emborough Quarry, Old Down, about six miles N.N.E.
of Wells (fig. l),he found a Triassic fissure deposit containing a great quantity of
well preserved but dissociated small reptile bones. I n the same year, at Batscombe,
near Cheddar (fig. l), he collected small reptilo bones, similar to those found at
Emborough Quarry, from another very fossiliferous Triassic fissure deposit. He also
found one small fragmentary reptile jaw in another Triassic fissure a t Highcroft
Quarry near Gurney Slade (fig. 1).
Kiihne next turned his attention to South Wales. The geological history of
Glamorgan in Triassic, Rhaetic and early Jurassic times is rather similar to that of
the Mendip Hills of Somerset. I n Glamorgan there were considerable outcrops of
Carboniferous Limestone, which remained partly above water as islands in Rhaetic
times, and were not submerged until the bucklandi zone of the Lower Lias (Trueman,
1922). Here he began to search the numerous quarries for fossiliferous &sure
deposits. Unknown to Kiihne, Mr. Trevor M. Thomas and his brother had discovered
a fossiliferousTriassic fissure deposit a t Ruthin Quarry, St. Mary Hill, near Cowbridge
(fig. 1). This locality has since been described by Mr. Thomas (1952), and a quantity
of the material collected has been presented to the National Museum of Wales,
Cardiff. Kiihne also found this locality and collected a small amount of material
which is now in the British Museum (Natural History). At Duchy Quarry, between
Southerndown Road Station and Ewenny, south of Bridgend (fig. l),Kiihne discovered
a small quantity of matrix which had been part of a fissure-filling. The fissure itself
262
PAMBLA LAMPLUGH ROBINSON : THE I\IESOZOlC FISSURES OF THE
had at that time been quamed away. The matrix yielded about fifty teeth, or
fragments of teeth. The fauria consists mainly of Triconodonts, almost identical
with those found by Professor Peyer in the Upper Rhaetic of Hallau. One tooth
proved to be that of a Symmetrodont, and is the oldest known record of this group
(Kiihne, 1949, 1950).
Dr. Kiihne returned to Germany in 1951. Continuing his work in the field,
I found a new Triassic fissure deposit a t Slickstones Quarry, near Cromhall in
Gloucestershire in 1951, from which a fauna of small reptiles has been collected
(Robinson, Kermack and Joysey, 1951). I n 1952 my colleague Dr.K. A. Kermack
joined in the hunt, and we discovered bone-rich material containing a t least five
different reptiles a t Pant-y-ffynon Quarry, near Cowbridge, Glamorgan, This
locality waa the Grst to yield associakl bones.
. . the desirability of strict
In his paper in 1946, Dr. Kiihne emphaaized
surveillance of all quarries in the Carboniferous Limestone of Somerset and South
Glamorgan. As quarrying continues, new fissures may be revealed at any time " .
These words have been fully confirmed by the experience of Dr. Kermack and myself.
I ' .
FIQ. l.-The
Mesozoic h u r e localities of the Brietol Channel ares.
The forreiliferous material from Pant-y-ffynon Quarry was found on the tip. No
trace remained of the iksure which had housed it. Another winter's exposure to the
weather would have resulted in complete disintegration of the material. Yet the
fissure had appeared Bince Kiihne had searched the quarry a few years previously.
In future all limestone quarriea in work in the Bristol Channel area will be inspected
annually, those in South Wales by Dr. Kermack, and those in Somerset and
Gloucesterak by myself.
In Table I all the fissurelocalities of the Bristol Chamiel area which have yielded
terrestrial vertebrates are listed, together with their contents as far as is known a t
present. Three l o d t i e a have yielded nine species of mammals or mammal-like
forms of Rhaetic or Liassic age. Six localities have yielded about nineteen different
apeoies of reptiles of Triassic age. The reptile fauna of Durdham Down, Bristol
(Theeodontcreaum antiquw Morris, T.Platyodon (Riley & Stutchburyj, and Pa2aeodaum c p k & m (&ley & Stutchbury) has been omitted from the list as the fossih
mem to have been found in a depreseion in the P h z o i a land surface and not in a
fissure (Riley & Stutchbury, 1840 ; Etheridge, 1870.
i
~ObbMtk!U8
abundant, fragmentary
A reptile (jaws)
Five reptiles
Ruthin Quarry, St, Mary Hill,
nr. Cowbridge, Clamorgan
Pant-y-ffynon Quarry. near
Bonvilston, Glemorgan
Triassic
Triassic
Triassic
Triassic
Triassic
1
Triasaic
Triassic
iRhtiC
Rheeto-Liassie
XLower Lies
I(Charmouthian)
* Generic name published as Clevosaurus due to a typographical error.
abundant
one fragment
A reptile jaw
Highcroft Quarry, nesr Gurney
Blade, Somerset
abundant
abundant specimens
a few specimens
rare
13 specimens
5 specimens
a few specimens
8 few SpeCh0XlS
The Emborough lepidosaur
2 other reptiles
About 9 species of reptiles
{ Qlevoaauncs h d m i Swinton*
Parrington
Microcleptid teeth
Incisom or cenines
Plecodont teeth '
Crocodilian teeth
EOZ08tPodon
Mb&ptea fnoorci (Owen)
Miorocleptes jEssurae Simpson
Thom&n avaglica Simpson
P Microcleptid incisors or canines
Reptiles
Eozostl.odonpaPanington
unidentified 6agments
1 specimen
1 specimen
1 specimen
about 50 apecimens
1 specimen
Other Tricononodont teeth
A Symmetrodont
Morganucodon walsoni Kiihne
about 2,000 specimens,
dissociated
Oligokyphw mjor Kiihne
{ Oligokyphw m i w Kiihne
Bataoombe Quamy, near
Cheddar, Somerset
Emborough Quarry, Old Down,
near Wells, Somerset
Slickatones Quarry, near
Cromhall, Gloucestershire
Holwell Southern (' Microlestes ') Quarry, Near
Frome, Somerset
Duchy Qu8rry, new Southerndown Road Station,
Glamorgen
Windsor Hill Quarry, near
Shepton Mallet, Somerset.
K. A. Kermack
1952 P. L. Robinson
{
T. M. Thomas
1946 W. G. Kiihne
1948 W. G. Kiihne
1946 W. G. Kiihne
1938 F. G. Hudson
1951 P. L. Robinson
1939 W. G. Kiihne
C. Moore
circa 1860
1947 W. G. Kiihne
1939 W. G. Kiihne
TERRESTRIAL VERTEBRATES FROM FISSURES I N THE BRISTOL CHANNEL AREA
TABLEI.
W
Q,
I
s
z
5
264
P A M E L LAMPLUQH ROBINSON : THE MESOZOIC FISSURES OF TEE
FISSURES.
The term fissure is a compendious one ; yet it is convenient to retain it t o refer
to a number of phenomena which have in common the fact that they consist of
relatively small quantities of younger deposits found enclosed in rocks of an earlier
age. In the Bristol Channel area there are several different kinds of fissure ; only
one of these, of submarine origin, has been described previously (Moore, 1867,
1881 ; Kiihne, 1946). This paper is mainly concerned with another type of fissure,
of terrestrial origin, in which pre-Rhaetic vertebrates have been found ; but a brief
description of three other types will be given.
Fissures arc small-scale phenomena, but their formation is invariably closely
connected with the general geological history of the areas in which they are found.
A brief description of the relevant part of the geological history of the Bristol Channel
area is a necessary preliminary t o a discussion of the fissures which occur there.
For the sake of brevity it is convenient t o select a part of the area, the Mendip Ells
of Somerset, for detailed examination and to serve, in general terms, as a n illustration
of the whole.
THE GEOLOGYOF THE UNDIP
AREA
The geological formations in the Mendip area are :r Great Oolite
Inferior Oolite
Jurassic
Lias
Rhaetic
I
MESOZOIC
1
\Triassic
Carboniferous
PALAEOZOIC
4
OF
SOMERSET.
Dolomitic conglomerate and red mads and
sandstones
Coal Measures
Grits
Limestone
BBISTOL OIULNNEL AREA AND THEIR VERTEBRATE FAVNAS
265
Lover Coal Sarles
FIG.P.-Diagram map of the four Mendip periclines showing the structure of the Palaeozoic rocks.
(After Moore and Trueman, 1939, by parmission.)
------__- _-- -- - * -
.u
FIG. 3.-The distribution of the facies of early Blue Lias in Somerset.
(From Kellaway 8nd Welch, 1948, by permission of H.M. Stationery Office.)
266
PAMELA JAMPLUQH BOBINSON : THE MESOZOIC FISSURES OF THE
stable than the depression south of the Mendip periclines. The relative stability
of the Codfield area compared with the southern depression was demonstrated again
in Lower Limic times (see fig. 3). The physiography of the Mendip area in Permian
and Triaesic times was therefore that of a ridge of high relief, controlled by the four
periclines, separating two areas of low ground to the north and south.
During Permian and Triassic times the high ground was eroded. The debris
was deposited to the north and south by intermittent torrents, as pebbly sands and
silt, and windblown dust was waterlaid in temporary shallow pools. Deposits of
gypawn, celeatine and rock salt indicate the presence of saline lakes in these a r m
towards the end of the Trias. Within the Triassic Mendip Hills screes of angular
rock debris, or more rounded pebbly torrent deposit, began to fill up the valleys
and accumulated on the hill slopes. The scree-torrent deposits are known as the
Dolomitic Conglomerate, and they, and the sands and marls, are often bright red,
due to the presence of finely disseminated haematite. The haematite may have been
derived from the pyritm present in Coal Measure shales, by weathering and oxidation.
The youngest Triassic deposits are often pale green, indicating a change in conditions
of deposition.
By the end of Triassic times, relief in the Mendip periclines had greatly diminished.
In the two eastern periclines over 11,OOO feet of Carboniferous rocks must have been
removed from the central parts of the folds, and in three of the four periclines a
central core of Upper Old Red Sandstone rocks was exposed. The most easterly
may
ill,have had the Old Red Sandstone core breached and
pericline, Beacon €€
Upper Silurian rocks exposed. The late Triassic relief achieved in the most westerly
pericline, Blackdown, contrasts with that of the other three. Part of it displays
inverted structural relief. The western part of its Old Red Sandstone core was
eroded to form a valley whose walls, formed of Carboniferous Limestone, were the
northern and southern limbs of the pericline. The central valley contains a filling
of Dolomitic Conglomerate and Red Marls, and the northern and southern limestone
walls were breached by other Triaasic valleys. The other three periclines had central
cores of Upper Old Red Sandstone or Silurian rocks which formed the highest ground,
and the structural relief thus remained uninverted.
At the close of the Triassic the Rhaetic Sea invaded Great Britain. Parts of t.he
Mendip periclines were not immediately submerged, some of the former high ground
remaining above sea level as islands, and the Rhaetic deposits become thinner and
more littoral in character as they approach these islands. The succeeding marine
deposits, Liassic, and Lower and Middle Inferior Oolite, show that the islands were
subject to encroachment by the sea, with intermittent periods of recession and
erosion of previously formed littoral sediments. It is probable that the islands were
completely submerged in Upper Inferior Oolite times, although positive evidence is
available only at the eastern end of the Mendips, where outcrops of Upper Inferior
Oolite have been preserved together with their relation to the underlying rocks.
The present condition of the Mendip Hills is due to the fact that post-Mesozoic
erosion has stripped off much of the softer covering of younger Mesozoic rocks, while
the more resistant PalaRozoic and Dolomitic Conglomerate rocks have been less
affected. Thia process of exhumation has revealed much of the older, Triassic,
topography. Today, as in Triassic times, the Mendips form a ridge of hilly ground
separating more low-lying areas to the north and south.
THE FISSURES
OF TRE MENDIP ABEA.
At least four different types of fissure occur in the area. Two of them originated
under the sea and are filled with marine deposits which may be Rhaetic to Inferior
Oolite in age, The other two are terrestrial phenomena and are Triassic in age. The
marine types will be discussed first.
The Neptunian Dyke. The characters of certain fissuressuggest that they originated
under the sea as open clefts or cracks in the older rock forming the sea-floor. In
the Mendip region the cleft ww usually an opened joint or bedding plane in the
BRISTOL OHAhTEL AREA AND THEIR VERTEBRATB FAUNAS
267
Carboniferous Limestone, which probably gaped as the result of small tensional
forces operating on the limestone in the area. The cleft was long in proportion
to its width, had sub-parallel walls, and a fairly stmight course. The long cleft
waa filled with the debris which shifted about on the floor of the Rhaetic or Jurassic
sea. This waa the type of &sure known to Charles Moore.
Two neptunim dykes were described by Kiihne. One waa exposed during
quarrying operations a t the Microhtes Quarry at Holwell,.and yielded all but one
of the mammalian teeth found by him (Kiihne, 1946). The other, still exposed
at Windsor Hill Quarry, Downaide, near Shepton Mallet, yielded abundant dissociated bones of OLigoky@~. Other dykes of this type occur in the easfern part
of the Mendips, near to the outcrop of Rhaetic and Jurassic rocks. Care must be
taken in estimating their age. The fillings may contain a mixture of pebbles with
fossils ranging from Rhaetic to Inferior Oolite in age, but only the youngest elements
indicate the time when infillina occurred. The dykes were situated close to the
Jurassic Mendip islands, and the derived material suggests that the marginal Jurassic
seas left thin deposits which were periodically reworked. The derived material
found in these dykes is of intereat, as it provides evidence bearing on the original
extent of former thin sedimentary covers which were removed during short periods
of Jurassic erosion.
As a possible source of Rhaetic or Jurassic terrestrial vertebrates the neptunian
dyke offers two possibilities. One is exemplified by the Holwell dykes, the other
by the dyke at Winhor Hill. I n the Holwell dykes the mammalian teeth were
almost certainly derived fossils of Rhaetic age redeposited in the fissure in Inferior
Oolite times. The labour involved in securing them waa considerable. The materid
sorted by hand by Charles Moore included about 70,000 fish teeth and yielded 29
mammalian teeth. Kiihne washed about 2,260kg. of clay from his ‘ Holwell2’
fissure and handpicked a concentrate of about 121 kg., which yielded 19 mammalian
teeth.
The Windsor Hill neptuni&n dyke presents a different yield of terrestrial
vertebrate bonea. In this case the bones are abundant over a limited part of the
fissure, and sparingly distributed over a rather greater part of its length. They
are all dissociated, but are evidently not derived, as they are hardly waterworn and
not limited to the most durable elements such as teeth. Kiihne suggested that
during Liassic (Charmouthian)tipes the &sure opened under the sea, close to the
shore of one of the Jurassic Mendip islands. Occasional heavy rains swept the
bones of Oligokyphus from the land surface into a brook, and so out to sea. The
fissure, being close to its mouth, trapped the heavier particles of the stream’s load,
including the bones. Thus rainwash and the stream were one factor in the areal
concentration of the bones, and the fissure waa a further factor. The state of
preservation of the bones suggests that the journey to their place of deposition was
short and that burial was rapid.
a gaping
Neptunian ‘ sagged-cover’ dykes. This type of fissure also originated a i ~
joint or bedding p h e of the P h z o i c rock which formed the Rhaetic or J u r w i c
sea-floor. But when the crack gaped a thin sedinentary cover already existed
above it, and this cover was let down into it. The cover deposits pa,rtially lined
the sides of the crack, but were also partidy disrupted to form clots and fragments
filling the centre. Angular fragments of the walls were often included in the fissure
filling. The rest of the filling consisted of contemporary marine deposits, sometimes
almost horizontally bedded, whioh constitute the matrix of the ewer fragmental
material.
example of this type of fissure occurs just over a mile north of Gurney Slade.
There me three exposuree of the same fiasure, which runs in an east-west direction.
a field on the west side of the road fkom Gurney Slade to Old Down is an old quarry.
A vertid wall of rock hss bean left u n q u d e d , as inferior stone, and stands isolated
on the grase-grown quamy floor. This is the western part of the &me. A steepsided outmop of Carboniferous Limestone, about 16 feet high, follows the east side
JOVBN. LIKN. 8OO.-ZOOLOGY,
VOL. XLIII.
19
26%
PAMELA LAMPLIT(:H ROBINSON : THE MESOZOIC FISSURES OF THE
of the road, and the fissure is exposed here iii cross-section. East of the road,
and behind this outcrop, is a disused quarry named Cockhill, in which the eastern
part of the fissure is exposed (Pl. 4, and fig. 4).
This fissure filling consists of Triassic and Liassic sediments. The Trias is usually
fine-grained and may be well-laminated ; a coarse conglomeratic facies is developed
higher up on the south side of the western exposure of the fissure. The Trias is
usually red, but occasionally a buff colour occurs. The Lias is most commonly a
grey, comminuted shell limestone, with small conglomerate pebbles rarely larger
than a quarter of an inch in diameter. Variations from this lithology occur, and a
fine-grained limestone crowded with well-preserved fossils, or a coarse conglomerate,
is sometimes developed. The conglomerate pebbles are usually quite well-rounded
and include quartz, chert, Carboniferous Limestone and occasionally Trias. The
Liassic fauna consists of lamellibranchs, gastropods, a rhynchonellid and an echinoid.
The fauna was submitted to Dr. L. R. Cox and Dr. H. Muir-Wood. Their identifications, tabulated in an appendix, suggest that the deposit is Lower Liassic, probably
Upper Hettangian. No evidence has been found to indicate that the Liassic
sediments vary in age in different parts of this fissure.
L E G E N D
1
s
N
c4t' d
RMm C
SECTION AT PostTuni A
c
A
D
7.
336
- i"
FIG.&--The western part of the Gurney Slade fissure. Lower diagram : The western part of
the fissure seen from the south ; Mesozoic rocks left blank.
The way in which the fissure and its filling originated is suggested by the character
of the walls, and the disposition of the Triassic and Liassic sediments. The Carboniferous Limestone walls of the fissure are not waterworn. They are usually vertical
and plane, but in placea form irregular projections. It seems unlikely that the
fissure was formed by solution of the limestone, but rather that it originated as a
cleft which gaped as the result of small tensional forces acting in the locality. The
Triassic fissure deposits occur chiefly as two vertical sheets, one lining each wall
of the fissure (fig. 4). Where a lamination is developed this is directed vertically.
The sheets of Trias are broken in places, and the broken edge may curl towards the
interior of the fissure filling so that the edge is embedded in Liassic sediment.
Detached fragments of the Triassic lining material lie in the more interior deposits
BBfSTOL aEANNEL AREA AND THEIR VERTEBRATE FAUNAS
269
of’ the fissure. Within the outer lining of Triassic material there may be a second,
inner, lining of Liassic sediments. I n places these Liassic sediments are finely
bedded, and the bedding dips vertically until it approaches the lower part of the
fissure, when the dip becomes almost horizontal and is directed towards the interior
of the fissure filling. Where the Triassic lining is absent, due to breakage, the
Liassic lining lies adjacent to the wall of the fissure. The more interior fissure filling
consists of Liassic material. This may exhibit an almost horizontal bedding, which
usually displays a sharp discontinuity with that of the sedimentary linings of the
fissure walls. Frequently the more interior Liassic fissure deposits show a rather
confused arrangement which suggests that dots of Liaasic material are embedded
in rather similar sediments. The distribution and arrangement of the Triassic and
Liassic sediments suggests that, prior t o the opening of the fisaure, a very thin
deposit of continental Trias, with small pockets of conglomerate, covered the Carboniferous Limestone in this locality. I n Upper Hettangian times a thin veneer of
marine deposits was added above the Triassic cover. The fissure then opened and
the two sedimentary covers sagged into it, partly lining the fissure walls and suffering
a partial disruption in the process. After the fissure opened the deposition of Upper
Hettangian sediments continued, formed a matrix to the fragmental material, and
accumulated until the fissure was filled.
No vertebrate fossils were found in the Gurney Slade-Cockhill fissure. The
possibility that, such fissures will yield terrestrial vertebrates is probably rather
less than in the case of the neptunian dyke. They were probably less efficient as
traps, because when formed they were immediately partly filled by the cover deposits.
If the cover deposits had included a bed containing terrestrial vertebrate remains
then these would be preserved in the fissure even if subsequent erosion had removed
them from their undisturbed position on the Palaeozoic rocks in the viciriity. As
in the case of the neptunian dyke, the possibility of finding terrestrial vertebrate
remains in this type of fissure depends to a considerable extent on the position of the
fissure in relation to the Jurassic Mendip islands, and to the agents of transport
obtaining at the time.
Subaerial &sure. This type of fissure is Triassic in age and terrestrial in origin.
Its character suggests that it originated as ascrack in the Carboniferous Limestone,
which opened on the Triassic land surface, and became filled with surface rubble.
An example can be seen a t Torhill Quarry, about half a mile east of Wells, on the
Wells-Shepton-Mallet road. The fissure has subparallel walls which show no sign
of water-wear. The Triassic filling is a breccia, composed of sharply angular fragments of Carboniferous Limestone, with a red matrix of more finely grained sediments.
The filling shows no trace of bedding and is completely unsorted. No fossils were
found in it.
Underground unztercourses. At the present day underground watercourses are in
process of formation in areas where hard, well-jointed and bedded limestones are
being subjected to the action of rainwater containing traces of carbonic acid. The
rainwater quickly finds its way down the joints and bedding plan-, and gradual
solution of the limestone walls along its route produces a system of underground
watercourses.
As soon as erosion had stripped the Coal Measure rocks from the Triassic Mendip
Hills the underlying Carboniferous Limestone was exposed to the action of rainwater.
Underground watercourses formed, and eventually filled with Triassic sediments ;
they are occasionally exposed in quarrying the Carboniferous Limestone. This
type of fissure can be recognized by the following characters.
The fissure filling is usually enclosed by a limestone roof, as well as limestone walls
and floor (fig. 6 ) . The limestone boundary walls are usually irregular in cross
section, in plan, and in profile. I n places they swell to form @wallcaverns (Pl. l),
in other places they form narrow irregular passages or mere cracks in the limestone.
The walls are u e u d y waterworn, and are often coated with stalactite. The sediments
of the fissure-filling are often well bedded (Pl. l), sometimes displa@g small
19*
270
PAMELA LAMPLUOH BOBINSON : THE MESOZOIC FISSURES OF TEE
S U L E IN FEET
FIG.&-The Sliokstonea Qusrry fissure.
s d e current bedding. The bedding is usually newly horizontal, although sometimes
oompaction of the sediments may have r d t e d in a considerable sag-curvature of
the bedding planea. The sediments are sometimes conglomeratic, the fragments
being Carboniferou~Limestone. Pieoes of stalactite, and recrystallized cave pea,&,
BBfiIBB a l U " E L AREA AND THEZa VERTEBRATE FAUNAS
271
are sometimes present in these
Ijllings. The best examples of this type of
fissure are exposed at Batscombe Quarry, near Cheddar, in the Mendip area ; and
at Slickstones Quarry, near Cromhall, in Gloucestemhire (see fig. 5 and P1. 1).
The only invertebrate fossils which have been found in these Triamic watercourses
are occasional specimens of Emtheria minutar (Alberti in Zieten) and derived Carboniferous Limestone forms. Not all of the Triassic underground watercourses,
which appear in the course of quarrying operations, contain terrestrial vertebrates.
But a high proportion of them contain a t least a few bones. These may be a few
poor fragments, sparsely distributed through a part of the fissure filling. But often
a real bone bed is present in a certain part, or parts, of the watercourse, always near
to the end at which water entered. The Triassic terrestrial vertebrates listed in
Table 1 have all come from underground watercourse deposits.
THE AQE OF TEIE 'TRIbssrc' FISSUREDEPOSITS.
The age of the underground watercourse sediments, and the bones which
they contain, is not easy to assess. They ase of continental origin and could therefore
have been deposited at any time during the Permo-Trim, indeed in a few limited
areas they could be Rhaetic or Jurassic. Their sedimentary fillings form isolated
pockets, and it has not been possible to trace a connection between them and the
main Permo-Triassic outcrop. Even if this could be done it would not help matters,
for, except at one or two localities, the Permo-Trias of the area is barren of foesils.
The vertebrates found in the watercourse deposits are new genera ; they belong to
sauropsid groups whose early Mesozoic history is often only sketchily known, and
there are few places in the world where a good sequence of fossiliferous horizons of
Permo-Triassic continental rocks have been found.
In the Bristol Channel area the earliest dated horizon of Mesozoic age is the base
of the Rhaetic. As conventional methods cannot be used, other means have had to
be employed in dating the continental underground watercourse sediments. Three
lines of approach to the problem have been found. A consideration of the probable
history of the formation and infilling of the Triawic underground watercourses allows
one line of evidence to be suggested regarding a probable age for their deposits. This
is applicable in a general way to all the watercourses in the Bri~tolChannel area.
A second line of evidence is supplied by the Emborough Quarry watercourse, and is
based on a detailed study of this locality. A third line of evidence is baaed on the
general character of the vertebrate fossila found in the fissure deposits.
The history of the underground watercourses may now be considered and the
Mendip area used w an illustration. The formation and infilling of the underground
watercourses is regarded as controlled by the relief and position of the water-table,
and the history of the water-table in respect of these two factors is closely connected
with the erosional and depositional history of the district in Permo-Triassic times.
At the sites of the cores of the Mendip periclines over 8,000feet of Coal Measures
and nearly 3,000 feet of Carboniferous Limestone were removed by erosion during
Permo-Triassic times. It is not necessary to postulate that the Permo-Triassic
Mendip Hills ever had the great relief which these figures suggest, as uplift and
posthumous movement of the periclines during the two periods may have rejuvenated
relief and kept active erosion in progress. Indeed the presence of narrow v d e y
heads, fled with Dolomitic Conglomerate, which may extend up into the Old Red
Sandstone core of some of the periclines, suggests that even at a late stage of general
denudation of the whole ridge youthful valley-forms existed, and mechanical erosion
waa active until very late in Permo-Triassictimes. The underground water channels,
formed after the Carboniferous Limestone had been exposed by denudation, imply
the existence of a certain amount of rainfall, and thus the presenoe of a water-table
in the area.
While Permo-Triassic topographic relief was still well defined and before rockwaste had banked up the slopes and filled the valleys and neighbouring lowlands, the
water-table mu& have had a fa& relief over the whole m a , and in the hills its level
2T2
PABfELA LAMPLUGH ROBINSON : THE MESOZOIC FISSURES OF THE
probably fluctuated considerably. Such conditions, promoting the flow of subsurface
waters, would be most favourable to the formation of underground water channels
in the hilly region. In the higher ground the level of the water-table must have fallen,
relative to rock horizon, as denudation end uplift proceeded. General lowering of
the surface, wbioh continued until late in the Triassic, must have removed much of
the earlier formed watercourses, and only the deeper parts of these, or those formed
late in the !Priaasic, can have been preserved. Little sedimentation can have taken
place in the underground water channels until the water-table reached a minimum
level towards the end of the Triassic. Thus the period of formation of the water
channels, when solution of limestone was the dominant process, rather than deposition
of sediments, continued from the time when the Carboniferous Limestone was exposed
until late in Triassic times.
Towards the end of the Triassic, as relief was planed down in the hilly area and
the neighbouring lowlands filled with sediments, the water-table must have become
more uniform in level over the whole district, and therefore relatively stagnant.
Within the hilly area, towards the close of the Triassic period, the U g of the
valleys may have produced a slight general rise of the water-table, owing t o the
disappearance of the small-scale relief features and the continued banking up of
the lowlands. These two factors, the greater stagnation of the water-table over the
whole district and the slight rise of the water-table within the hilly area, led t o the
silting up of the underground watercourses, for run-off waters entering solution
channels filled with water would soon have their velocity checked and deposit their
load. It is not suggested that the silting up of the underground watercourses
occurred synchronously over the whole of the Bristol Channel area or even in the
Mendips-indeed a sequence of iufilhng would be expected. But the period in
which deposition was the dominant process in the underground water channels was
of short duration relative t o the length of Permo-Triassic time, and it occurred late,
just prior to the Rhaetic invasion. As the sea did not cover the entire Mendip area
until Middle Jurassic times, nor quite submerge the Carboniferous Limestone outcrops
of Ghmorgan until Lower Liassic times, there remains the possibility that some of
the watercourses continued t o fill with terrestrial sediments after the Rhaetic invasion.
hi all the fossiliferous underground watercourses discussed in this paper, and listed
as Triassic in Table I , except that a t Highcroft Quarry, marine Rhaetic can be shown
to have covered the site of the fissure. The underground watercourse deposits of the
Bristol Channel area and the terrestrial vertebrates which they contain, are therefore
probably Upper Triassic in age.
A study of the Emborough locality provides evidence regarding the age of ith
watercourse sediments. Emborough Quarries are situated on each side of the
railway line running from Radstock to Shepton Mallet (see fig. 1 and PI. 4).
The entrance to the Quarries is on the Radstock-Wells road between Emborough and
Old Down, about six miles north-east of Wells. The Triassic deposit in which the
reptilian bones were found is situated in the eastern quarry. The limestone face
has been worked back on each side of the Triassic deposit, so that the latter, left
rejected
inferior stone, forms a projection which juts out from the eastern face.
The end of this Triassic projection faces approximately north-west, and the two
sides face approximately north-east and south-west (Pl. 4).
The south-west side consists nia.inly of Carboniferous Limestone which has the
same dip and strike as that of the adjacent quarry face. It represents one of the
lateral walls of the watercourse. I n its lower part some of the joints and bedding
planes are waterworn, with small swirl-holes filled with finely bedded red clay.
The Carboniferous Limestone on this south-west side forms a very thin flankmg wall
left in contact with the watercourse deposits, and most of the projection consists
of a mms of Triassic sediments (Pl. 3).
The Triassic sedinients can best be seen on the north-west end of the projection
(PI. 2). The lower part of the mass consists of a homogeneous, well-bedded,
lnicaceous clay, dark red in colour but with occasional streaks of green. The upper
BRISTOL CHANNEL AREA AND THEIR VERTEBRATE FAUNAS
273
part is a conglomerate in which large boulders of Carboniferous Limestone, up to
several tons in weight, are set in a matrix of limestone pebbles, flakes of a pale shale,
and silt. I n places, where the silt is pebble-free, it is finely bedded. The silt is
usually deep red, but occasionally a pale green colour is developed, either as patches
or as alternate laminae in the bedded silt. Pieces of stalactite have been found in
this conglomerate. The large boulders are usually waterworn on one or two of their
faces, the others being angular fracture surfaces. Both the worn and the angular
surfaces lie embedded in the mass of pebbles and silt, showing that the condition of
the boulders was produced before the deposition of the surrounding sediments.
There is an abrupt change in deposit from the lower homogeneous clay to the
conglomerate above.
The north-east side of the Triassic projection has been much obscured by the
tipping of quarry rubbish. The visible sediments are all of the conglomeratic type.
The lateral wall of limestone which should lie in contact with the Triassic sediments
on this side has been quarried away. The fossil bones occur towards the top of the
upper, conglomeratic sediments, on the north-west and north-east side of the
projection .
The whole Triassic deposit is interpreted as the sediments of a cave, forming
part of an underground watercourse. At a certain stage in its history the cave roof
and part of the sides collapsed. The homogeneous, well-laminated clay, which
forms the lower part of the depoait, represents the fine sediments laid down in the cave
before the collapse occurred. It suggests that the coarser and heavier part of the
stream’s load had been dropped before this cave was reached. At this stage, therefore,
the cave may have been some distance from the intake end of the watercourse.
The conglomerate, which begins abruptly above the clay, represents the collapsed
roof and upper sides of the cave. When collapse occurred the cave would be in
direct communication with the land surface, having become a depression filled with
limestone blocks. The silting up of this depression would be partly due to the
stream running through the bottom of it, but mainly due to the trapping of rainwash
from the surrounding land surface. The fossil bones were found towards the top
of the conglomerate ; they are not found in the lower part, nor in the clay deposited
prior to collapse. It is probable that they were swept into the depression, with
other debris, from the surrounding land surface after periods of rain.
The presence of such a large number of limestone blocks in the conglomerate
suggests a total collapse of the cave roof, as distinct from the dropping of a few
unstable blocks. This indicates that just before the collapse occurred the cave roof
had become quite thin, the contemporary land surface lying perhaps only a few feet
above. If the age of this land surface could be estimated it would provide an approximate date for the collapse of the cave. The age of the debris and reptile bones filling
the collapse depression would then be post-collapse and pre-Rhaetic.
I n order to form an estimate of the age of this Triassic land surface it is necessary
to study the Mesozoic topography developed on the Palaeoeoic rocks of the area
round Emborough Quarries. The Palaeozoic rocks of this area form the eastern part
of the Pen Hill pericline (fig. 2). In the discussion which follows it should be borne
in mind that the Mendip Hills were not a glaciated area in Pleistocene times, but
were subjected to periglacial conditions. In the Emborough area the effects of the
periglacial climate are not very striking and were possibly limited to the break-up
of part of the thin Jurassic cover by frost-action, followed by incorporation of the
resulting blocks into a subsoil rendered pasty by thaw.
The Mesozoic topography falls into two categories. A Triassic topography of
hill and valley showing a fair degree of small scale relief has been buried and to some
extent re-exposed. This topography can be shown by mapping the distribution
of the Dolomitic Conglomerate. The map (Pl. 4) shows that the Conglomerate
fills a, s m d narrow valley to the south and south-west of the Quarries, and the
outcrop then curves northwards and finally westwards. I n its northward and
westward c o m e the Conglomerate outcrop is following round the curve of a hill of
274
PAMELA LAMPLUOR ROBINSON : THE MESOZOIC FISSURES OF
Palaeozoic rocks and is banked against its slope. The spur of the hill coincides
with the ’ nose ’ of the Pen Elill pericline ; the Triassic valley was eroded in the
zones of the Carboniferous Limestone which form the Binegar syncline. Emborough
Quarries, and the Triassic cave, lie on the spur.
The other element of Mesozoic topography is more closely related to the present
surface and is shown on the map (Pl. 4). The land surface slopes gently from
about 800 feet in the southern part of the map to about 700 feet in the northern,
and passes across the Palaeozoic rocks and the Dolomitic Conglomerate. Quaternary
and recent erosion has begun to dissect this surface by cutting down from it a series
of narrow winding valleys. The land surface passes across the top of the Emborough
Quarry faces and also across the Triassic cave-deposit. To the north of the Quarries,
it is cut by a narrow valley, but can be seen again further north where the Quarry
officer, and plant are situated.
This surface, which slopes gently northwards, is that on which the Rhaetic and
Jurassic sediments w-ere laid down, for small isolated patches of these rocks still lie
on it in places. At the cntrancc to Emborough Quarries, at the northern edge of the
map, part of the ruain outcrop of the Rhaetic appears. Here it is only 15 feet thick,
and consists mainly of black shales, pale clays, and grey sandstones which weather
yellow (Morgan & Reynolds, 1899). This outcrop is now overgrown. Four
hundred yards south of the Rhaetic outcrop, at the top of the face in the eastern
quarry, where the land surface is present, occasional fragments of black shale and
blocks of a yellow sandstone containing plant fragments and casts of small lamellibranchs can be found in the subsoil above the Carboniferous Limestone. There is
little doubt that these are remain4 fkagments of a former Rhaetic cover, leached by
subsoil weathering. The lithology of the sanhtone is quite different from that of
the later Jurassic rocks in this area. South of Emborough Quarries the neptunhn
mgged cover dyke, described in a preceding section, preserves a sample of a former
I h a o i c cover which has since been stripped off a t this locality. No Rhaetic could
be detected in the dyke. It is possible that a Rhaetic cover existed here and was
stripped off in pre-liassic times, but this seems unlikely as the thin, often finely
leminated, Triassic cover would also have been removed. The Rhaetic which occum
a t the entrance to Emborough Quarries is thin, and it seems more probable that the
Rheietic feathered out between Emborough Quarries and the dyke. South-east of
Gurney Slade occurs a silicified sandy rock marked on the old Survey Map (O.S.1 inch,
sheet 19), as Lias, but the contained fossils (see Appendix 11) prove that this rock is
actudly Inferor Oolite.
If this land surface represents the plane of the Rhaetic-Jurassic transgressions,
then it also represents, in an approximate way, the final surface of relief achieved
here a t the end of the Triassic, slightly trimmed by the invading seaa. From its
present position in relation t o the Emborough cave deposit there is little doubt that
it must lie only a few feet below the Triassic land surface which existed hereabouts
a t the time when the cave collapsed. The Triassic collapse-surface must therefore
represent a very late stage in general denudation of the area ; the last few feet of
more than 8,000 feet of rock removed from this part of the pericline. The debris
and bones which 6lled the collapse-depression in this surface must therefore have
been deposited late in the Triassic period.
The character of the vertebrate fauna obtained from the Triassic fissures of the
Bristol Channel area provides some evidence of its age. Von Huene (1940), who
utudied the Permian and Triassic faunas of the ‘ Gondwanaland’ areas, the United
States, Scotland, and Rusgia, has shown that during the Middle Triassic a marked
faunal change occurred. In the earlier part of the Triassic the mammal-like reptiles
are the commonest faunal elements: diapsids axe relatively rare. In the Upper
T r i d c the diapsids are the common forms, the archosaurs in particular having
begun a radiation into a number of orders and families, and the mammal-like
reptiles have become extremely rare and ‘ advanced * in structure. No mammal-like
reptib have been found in the Triassic watercourse deposits to date. The fauna
wnSiFlts entirely of diasds and pmtorwaurs.
BRISTOL CHANNEL AkEA AND THEIR VERTEBRA^ FATJNAS
275
Thus the Triaasic &sure faunas are, in general chwacter, Upper Triassic in type.
The geological evidence suggests that they are late Upper Triassic in age, both on
the general hypothesis discussed above, and on a different line of evidence obtained
from the study of a single locality. This suggests that the faunas from each locality
may possibly form part of a larger Bristol Channel fauna of late Tria&sic age. The
distribution of the members of the Bristol Channel fauna adds a little weight to this
last suggestion, although the areal distribution of reptiles is stilI inadequately known.
The reptile commonest in the Emborough fissure deposit also occurs a t Batscombe
Quarry, Cheddar, and is present aa a rarity in the 1951 Slickstones material. This
Slickstones deposit also contains a small reptile, probably a lepidosaurian, which
has a very characteristic dentary. Several specimens of an exactly 8imila.r dentary
have been found in the Pant-y-ffjmon fauna, from near Cowbridge. The fact that
two reptiles occur at more than one locality is suggestive. If this were the case with
one form only it might be an ubiquitous species with a fairly long range in time.
It is hoped that further preparation of material will reveal other forms which allow
cross-correlation between localities.
The available evidence for age, whether supplied by the vertebrate fauna or from
geological investigation, is corroborative. It suggests that the watercourse deposits
discovered to date are late Triassic, and the terrestrial vertebrates which they
contain form a Bristol Channel fauna of this age. It is hoped that the geological
evidence of the age of the Bristol Channel fauna, relative to the Rhaetic invasion of
western Britain, may be helpful in assessing correlation between the horizon of its fauna
and horizons of continental vertabrtaas in other areas.
TLandsurface
time of collapse
at
-
Position ofpresent
Landsurface
CARBONIFEROUS
LIMESTONE
I TRIASSIC. COLLAPSE OF CAVE
,RhaeLic -Jurassic
Tran~gre=sionPlane
J U R A S S I C --A
RHAETIC -CARBONIFEROUS
LIMESTONE
II. MID -JURASSIC
Site ot
hternarv
Triassic
Rhoelic
-
Jurasslc
outlier
Present Landsurface
wprox = RhaeticJumssic Transgression
Plane
RHAETIC
CARBONIFEROUS
LIMESTONE
-
IU. PRESENT
FIQ.B.-The
DAY
history of the Emborough Quarry fissure.
LATETRIASSIC CLIMATEAND VEGETATION
IN THE BRISTOL
CHANNEL AREA.
A Characteristic feature of the Triassic rocks in the area is the prevailing red
colour. This red colour is usually absent towards the top of the Trias, and may
be absent l o d y at any p& of the period. Although the mode of formation of
such red be& is a subject of controversy no one has yet suggested that extensive
deposits of them were formed under temperate, cold wet, or cold arid conditions.
The Olimstio oonditions which have been suggested as necessary for their formation
are either hot md dry,or hot and humid.
‘376
PAMELA LdMPLUGH ROBINSON : TIIE MESOZOIC FISSURES OF THE
Under hot humid conditions chemical weathering plays the more active role in
rock destruction, and under its influence only the most stable minerals are preserved.
Fragments of Old Red Sandstone, obtained from the Dolomitic Conglomerate, show
no signs of the effects of chemical weathering and still retain occasional chips of fresh
iuicrocline felspar. An analysis was made of the sediments in the Emborough
Quarry &sure. 92 per cent of the light mineral fraction consisted of flakes of felspar
The felspar may be partly derived from the Old Red Sandstone, but the noncalcareous residue of one of the types of limestone present in the eastern quarry
a t Emborough consisted mainly of cleavage flakes of felspar. The survival of such
a quantity of felspar flakes in the cave sediments (in both pre- and post-collapse
deposits) suggests either that hot and humid conditions prevailed but that erosion
and deposition was so rapid that the felspar flakes had no chance of decaying before
burial, or that deposition was at a moderate rate under semi-arid or arid conditions
with chemical weathering a t a minimum. Evidence which suggests that deposition
in the underground watercourses was not a t all rapid will be cited below.
Under arid or semi-arid conditions mechanical weathering is the more active
agent of rock destruction and chemical weathering is slight. The spreads of Dolomitic Conglomerate, containing enormous numbers of fragments of Old Red Sandstone
and Carboniferous Limestone, suggest that mechanical shattering, under arid or
semi-arid conditions, was a very active process in Triassic times.
Evaporite deposits occur in the Upper Triassic of the lowland areas, to the S. of
the Mendip range, to the N. of the Mendips and in S. Gloucestershire, and to a
lesser degree in S. Wales. Their presence suggests a hot climate in Upper Triassic
times, with a high rate of evaporation of saline lakes. Evaporites also suggest a
low rainfall, for abundant rain would provide a copious supply of fresh waters entering
the lakes and so check the effectasof evaporation by continual dilution.
The presence of the underground watercourses, channels dissolved in the Carboniferous Limestone and filled with water-laid sediments, requires that a certain amount
of rainfall should have occurred in Triassic times, a t least in the hilly areas where the
watercoums were situated. I n view of the evidence for a semi-arid or arid climate,
just reviewed, it is reasonable to postulate an intermittent rainfall which may have
been moderately heavy but which occurred a t infrequent intervals. The sediments
in the underground watercourses provide some evidence in support of this suggestion.
Most commonly they are well-laminated deposits (see PI. l ) , and mud cracks
frequently occur in them, indicating that the watercourses often dried up. The
mud cracks, which may be up to an inch across, are sometimes filled with small
chips of limestone, fragments of calcite and crinoid ossicles , and similar material
forms a thin layer through the channel desposit in continuity with the top of the
mud crack. These mud crack fillings, markedly coarser than the average deposits
of the channels, suggest the sudden onset of heavy rain which first brought in a load
of coarser debris from the land surface above before contributing fine silt and clay.
The presence of Euestheriu minirtu in t,he underground watercourses of Pant-yEynon and Slickstones Quarry. and its mode of occurrence in the main Upper Triassic
outcrop, is also suggestive of semi-arid conditions with an intermittent rainfall.
Present-day species of Estheria occur in temporary pools in slight depressions or
’ vleys ’ in the sandy plains of S. Africa.
Intermittent rains carry mud, dissolved
halts, and fragments of the sparse desert flora into the pools : EstlLeria appears,
flourishes, and then becomes buried in the mud as the pools dry out. The Upper
Triassic Euestheriu minuta, often occur in miall lenticles of silt, sometimes containing
salt pseudomorphs, suggestive of conditions similar to those seen today in S. Africa
(see Neaverson, 1955).
Some vegetation must have existed in the Bristol Channel area in late Triassic
times t o support the fauna of reptiles which lived there. No trace of it has so far
been found in the form of macrofossils, though the watercourse desposits still have
to be examined for the possible presence of spores. A small flora was discovered a t
Bromsgrove in the Midlands, but i t is earlier in age than the underground watercourse
sediments (probably Upper Bunter, see Wills, 1948, p. 85). It is unlikely that the
BRISTOL U U N E L ARXA AND THEIR VERTEBRATE FAUNAS
277
vegetation which eHisted in the Bristol Channel area formed a close cover, as this
would have retarded the mechanical weathering, whose importance has already
been mentioned. A rather sparse vegetation best accords with the climatic conditions
outlined above.
THEACCUMULATION
OF THE BONES
IN THE UPPER
TRIASSIC UNDERGROURD
WATERCOURSES.
The fossil bones found in many Tertiary and Quaternmy caves occur either in a
cave-earth or in a drip-tufa deposit, usually in a cave near to the outlet end of the
underground watercourse system. I n such cases most of the bones were brought
into the cave by animal locomotion, representing the remains of large and small
animals which used the cave for shelter and also those of the prey of the carnivores.
The bones accumulated in the Triassic underground watercourses in quite a
different way. They occur in water-laid deposits, always near to the inlet end
of the watercourse system, and were brought in by stream-water from the surrounding land surface. Commonly the bones are well-preserved, and must have been
buried quickly under a layer of silt. But sometimes the bones have been knocked
about, broken, and even rounded up into bone-pebbles before being buried. Occasionally associated remains occur, skeletons or parts of skeletons. At the present
day, in hot and fairly dry climates, the corpses of small terrestrial vertebrates may
dry up quickly after death, becoming very light ‘ mummies ’ of bones and dry skin
and flesh. If this also occurred under the semi-arid conditions of Upper Triassic
times the light mummies of small dead reptiles could, in some cases, be washed down
into the underground watercourses a t the onset of a heavy rainstorm.* Once
buried below ground under a layer of wet silt the skin and flesh would rot, but the
skeleton might remain intact unless disturbed by the next wash-through of rainwater. Locally dissociated reptile bones are sufficiently numerous to form a bonebed. This concentration of animal remains may be explained if the conditions on
the land surface in Upper Triassic times are borne in mind. I n the absence of a
close vegetation cover a rainstorm might sweep the debris from a fairly wide area
of land surface into the relatively restricted area of the nearest underground streanichannel, so effecting the concentration of animal remains.
!THE UPPERTRIASSIC
VERTEBRATE
FAUNA.
One of the important featur& of the Upper Triassic Bristol Channel fauna is that
its members lived on the higher ground of the continent of that period ; a feature
which emerges when the localities a t which they have been found are plotted on a
reconstruction of the relief in the area in Upper Triassic times. On high ground
erosion is the dominant process, and therefore little is known of the faunas which
have lived in such areas in the past, a t least prior to Tertiary times. Most terrestrial
vertebrate faunm of any period are lowland or coastal forms, for these are the
main areas where deposition of sediment occurs. W. D. Matthew, when discussing
dinosaur evolution in his book ‘ Climate and Evolution ’ (1939) wrote (p. 12) :
“The dominant order of land reptiles up to the close of the Mesozoic was the
dinosaurs, pre-eminently a dry-land adaptation in their inception, since their most
marked characteristic lies in their long limbs, bipedal progression and general
parallelism in proportions and structure to the large ground-birds of modern times,
which are today peculiarly the inhabitants of arid regions. The relationship and
origin of the more specialized, mostly gigantic, dinosaurs of the later Mesozoic can
best be explained by regarding them as a succession of derivatives from smaller
and more lightly constructed upland dinosaurs, mostly unknown to us, the larger
and more specialized types being re-adapted to a swamp life and inhabiting the coast
mamhea whose sediments are still preserved, while the more direct line of dinosaurian
evolution inhabited the uplands, where the sediments, if such were deposited, have
long since been removed by erosion, and the fauna is consequently unknown to us,
* A small reptile akaleton with part of the skin preserved has now (1955) been found in the
Sliokabnes Querry b u r e deposit.
2%
PAMELA LAMPLUGE ROBmSOH : TBE MESOZOIC P'lSSURl8S OF THE
except by inference . . . if our knowledge of the Tertiary sediments were limited
to the coastal swamp desposits, if in this country for instance we knew only the Tertiary of the Atlantic and Gulf coasta, we would be equally a t a loss for any direct
ancestral series illustrating the evolution of the mammalia ". Matthew contrasts
the record of the evolution of other terrestrial reptiles with the more adequate one of
many of the swamp-dwelling, amphibious or littoral forms. The Upper Triassic
reptiles found in the underground watercourse deposits of the Bristol Channel area,
which lived on the denuded higher ground of that part of the North Atlantic continent,
provide us with the only record known from Mesozoic rocks of a fauna of this
character. When the members of this fauna have been described it will be
interesting to compare them with faunas of the same period from the lowland areas
of the North Atlantic continent.
Another feature of the whole fauna is the size of its members. The largest reptile
present probably had an average size of about six feet. The smallest reptiles
are as yet inadequately known, being completely dissociated, but their dentaries
may be used as an indication of their size. Some of the small dentaries seem
likely to belong to individuals representing growth stages of rather larger forms.
Others, already known from a number of specimens, present a much narrower range
in size and are all small. The smallest dentary is about 8 mm. long, and has a tooth
row 4-5 mm. long bearing 12 teeth. Evidently some of the reptiles in this fauna were
about the size of a modern wall lizard (Lacerta muralis). In the Triassic the diapsid
reptiles began a radiation into the various orders which form the common elements
of the Jurassic and Cretaceous faunas. An Upper Triassic fauna of small diapsids
may be very valuable, as the small forms of any stock are held to be those with
evolutionary potentialities, being free from the specializations which are the
concomitant of large size.
In view of the age of this fauna and the size of its members the absenceof
mammal-like reptiles or mammals is particularly interesting. After the incursion
of the Rhaetic-Liassic seas immediate evidence of such forms is sound. Possibly
their ecological requirements were such that they moved northward, just in advance
of the Rhaetic-Liassic sea which brought a change of climatic and other conditions
to the areas on which it encroached.
British Trimsic rocks are notoriously unfossiliferous. Occasionally a small
fauna or flora is preserved. Such are the well-known Elgin reptile fauna, the small
flora and fauna from the Bromsgrove area of the Midlands,and the Durdham Down
reptiles from near Bristol. Elsewhere the usual record is a very occasional fragment.
These scattered finds give the impression that the British part of the Triassic continent
was not an area inimical to life, but rather that most of the life that existed was
not preserved. The Upper Triassic watercourse fauna heightens the impression that
an abundant and varied fauna existed on the continent of the period. The fomiliferous material from Emborough, Batscombe and Slickstones is so rich in bones that
in some places it forms a bone bed. The preservation of the material is excellent,
particularly at Emborough and Slickstones. That so little of the life of the Triassic
period has been preserved in Britain is no doubt due to a number of factors. One
of these factors may have been that the animal and plant remains were not buried
quickly but were subject to the slow process of comminution at the surface. The
underground watercourses probably afforded the only possibility of quick burial,
and subaequent protection, of animal remains in their vicinity.
The underground watercourses of late Triassic times are thus of importance as a
means of preservation of fossil bones and as providing a sample of a fauna from areas
of relatively high ground at a time when the evolution of the sauropsid reptiles was
at a particularly interesting stage, particularly in the case of the smaller representatives of the various groups.
.
ACKNOWLEDQMENTS
I should like to thank Professor S. E. Hollingworth €or reading and criticizing
the manuscript of this paper, and for discussing some of the fissures in the field.
BRISTOL CHANNEL AREA AND THEIR VERTEBRATE FAUNAS
279
I am grateful to Profeeaor D. M. S. Watson, F.R.S., for helpful discussions ; and
to the Head of my Department, Professor P. B. Medawar, F.R.S., for reading the
paper in mmusoript. To the Geological Survey officers, particularly Dr. Welch
and M i . Kellaway, I owe thanks for help and useful discussions ; also to Mr. Melville,
Dr. L. R. Cox, and Miss M i r Wood for identifying the invertebrate fossils listed in
the appendices. Dr. J. G. Capewell very kindly checked the identity of the minerals
present in the Emborough fissure-deposits.
SmMARY.
Terrestrial vertebrates have been discovered in Mesozoic deposits occurring &s
fissure fillings in the Carboniferous Limestone of some of the counties round the
Bristol Channel (Somerset, South aloucestershire, Glamorgan). These terrestrial
vertebrates fall naturally into two groups : one group consists mainly of mammals
and mammal-like reptiles, of Rhaetic or Lower Liassic age, which am found in
fissures of submarine origin. Most of these fossils have already been described,
and a summary ie given of Kiihe's studies of the fissurw in which they are found.
The other group consists of sauropsid reptiles only, found in a different type of
fissure, of continental origin.
The paper is mainly concerned with the second group of reptiles and the fissures
in which they am found. These Gssures we shown to be small underground watercourses, formed during Permo-Triwsic times when the area was part of a continent.
The age of the sediments and reptile bones found in these underground watercourses
is considered, and shown to be late !Massic, both on geological grounds and on the
character of the fossil reptilea. The reptiles from Werent localities probably form
part of a fauna, the Bristol Channel fauna, of late Triassic age.
In considering the age of the sauropsid reptiles a reconstruction of their physical
background is outlined, and this is completed by a tentative sketch of climatic
conditions. From this reconstruction emerges an important feature which characterizes the whole Bristol Channel fauna, namely, that the members of this fauna
lived on the higher ground of. the continent of the period. This is the only known
example of a mwozoio fauna of this type, and when its members have been studied
it will be interesting to oompme them with those from the lowlands of the Upper
Triassic continents.
The members of the Bristol Channel fauna identified so far are all diapsids and
protorosaurs. During the Triassic the diapsids radiated into the orders and suborders whose members form the common elements of later Mesozoic terreatrial
faunas. These late Triassic diapsids are therefore particularly interesting, especially
as those of the Bristol Channel fauna are all small (lees than six inches to over six feet
in length) and therefore fiee from the specializations which are the concomitant of
large size.
REFERENCES.
On
the
geologioel
position and geogrephicel distribution of the Reptilien
ETEEBXLK+E,
R. 1870.
or Dolomitic Conglomermte of the Bri~tolares. Quart. J . @sol. Soc. Lond., a6, 174-192.
H O ~ ~ C + W O B8.T E..
E , TAYLOE,J. H. &, KJZILAWAY,
Q. A. 1944. Lmge scale mper5ciel
S t N O t U m s in the Northempton Ironetone field. Ib&%. 100, 144.
v. HUENE,F. 1840. Die Seurier der Kerroo-, aondwene-,und verwandtan Albegeruugem in
fami~tischer,biologiecher und phylogenetisoher Hinsicht. Neues Jb. Miner. Xh. (B),
83,248347.
Golosy :B W and &uccsts,
KELLAWAY,
Q. A. & WEME, F. B. A. 1948. Bfitwh RegD4ettiot. GeoL Sum., London.
' Holwd 2 ' ; the ege-detarmhation
K U m , W. G. 1946. The geology of the fiesure-of the menun& '(u1teeth therein ; and 8 report on the teohnique employed when cpolleoting
the teeth of Eozeak.odon end Miaroleptidee. Proo. ~ 0 0 1 . Soc. Limd., 116, pte. XII & IV,
72833.
l
in the
1949. Exhibition of npechem from lodtiea of Meeosoic h ~ e s t r i e vertebretee
Briatol Chennel arm. Pm. Qkol. Soc. M.,
no. 1448.
1949. On a Trioonodont tooth of e new pettam from 8 fissure-fillingin kuth Qlemorgen.
Pmc. zool Boo, Imnd., IiB, pt. It, 34860.
---,
1960. A 8ymm6trodont tooth from the Rhwto-Liae, Patwe M.,166, 096,
-.
-.
580
PAMELA LAMPLUGH ROBTNSON : THE MESOZOIC FISSURES OF THE
M\~uML;R~-RIE,
J . IS%
Notes on an Autuirm excursion on the Mendips.
Proc. Buth Nut. Hi&.
~9oc., 5, 98-Ill.
A I A W.
~ ~
D. ~ 1939.
~ , C'liinate and Gvolut,ion. S p . . Pub. N.Y. A d . SCi., I, i-xii, 1-223.
MOORE, C, 1867. On abnormal conditions of secondary deposits when connected with the
Somersetshire and South Wales Coal-Basin ; and on the age of the Sutton and Southerndown Series. Qicnrt. J . Geol., Soc. Lotad., 23, 449-568.
-.
1881. On abnormal deposits in t.he Bristol dktlict. Ibid. 37, 67-82.
MOORE, L. R. & TRUEMAN,
A, E. 1939. The Structure of the Bristol and Somerset coalfields.
Proc. Ueol. Aw., 50, 16-67.
M o ~ o a C.
~ , I,. & REYNOLDS,8. II. lS99. Triassic deposits at Emborough. Proo. BriHd
Nut. SOC. (3), 9, pt. 2, 109-117.
NBAVERSON,
E. 1955. Stratigraphical Palaeontology. Ciarendon Press, Oxford.
OWEN, R. 1871. Nonograph of the mammalie of the Mesozoic formations. Palaeocatoqr. Soc.,
24.
PARRINGTON,
F. R.. 1941. On two inaminalian teeth from t,he Lower Rhaetic of Somerset.
Ann. X a g . Nut, Hist. ( l l ) ,8, 110-144.
__ 1946. On a collection of Rhaetic mammalian t,eeth. Proc. -001. SOC.Load., 116, pts.
111 and IV, 707-28.
RICHARD~ON,
L. 1911. The Rhaetic and contiguous deposits of tVest, Mid. and parts of East
Somerset. Qqmrt. J . Ceol. Soc. Lond., 67, 1-74.
RULEY,H. & STZ~TCHBURY,
S. 1840. A description of various fossil remains of three distinct
Saurian animala recent.ly discovered in the mapesian conglomerate near Briatol. Trans.
Geol. Soc. (2), 5, pt. 2.
ROBINSON,
P. L., KEBrdACK, K. A. & JOYSEY,
K. S. 1925. Exhibition of a new Upper Triaseio
lend fauna from Slickst,ones Quarry, Gloucest,ershire. Proc. Geol. SOC.Lorad., no. 1485,
.
86-87.
SIMPSON,G. G. 1928. A catalogue of tlg Memzoac I l l a n i d t b in the Geological D e p w t m t
of the British Museum. British Museum (Natural History).
XWINTON, W. E. 1939. A new Triassic Rhyncocephalian from Gloucestershire. Ann. Mag.
Nat. Ha. (?I, 4, 591-94.
THOMAS,
T. M. 1952. Notea on the structure of some minor outlying occurrences of littoral
Trise in the Vale of Ulemorgan. #eoZ. Mag., 89, 153-62.
TRUEMAN,
A. E. 1922. The Liaseic rocks of Glamorgan. Proc. h l . A88., 33, 245-84.
WELCH,
F. B. A. 1929. The geological stmxture of the Central Mendips. Quart. J . Geol.
Soc. W.,
85, 45-76.
. 1933. The geological structure of the Eastern Mendips. Ibid. 89, 14-62.
WILLS,L.J. 1910. On the fossilifemus Lower Keuper rocks of Worcesterehire, with descriptions
of some of the plsnta and animals discovered therein. Proc. Geeol. Ass., 21, 249-331.
. 1948. The palaeogeography of the AfiifAzno%. Univeraity Press, Liverpool.
WOODWABD,
H. B. 1876. Geology of East Somerset and the Bristol coalfields. Mem. ( f e o l .
-
-
Suw., England and Wales.
APPENDIX I.
FOSSILSOBTAINED OM THE GURNEY SLADEFISSURE.
BRACHIOPODA
Hettangian and basal Sinemurian.
Pdcirhynchia &ria
S.S. Buckman
LAMELUBRANCHIA
Chiracteristic of the angulatum.zone.
Modiolus hillanoides Chapuis and
Dewalque
Planorbis to bucklandi zones.
iWyowncha psilonoti Quenstedt
Cheksmys te&ria (Scblotheim)
Long-ranging Liassic species.
Lam (PlagiO8tomQ)sp.
A relatively small form of the group
L. gigantea ( J . Sow.). Liassic.
Pseudolimea pectin&&
( J . Sowerby)
Long-ranging Liassic species.
GASTROPODA
Pleurdoplaaria cf. conceuata
Terquem and Piette
pleu-e
sp.
Diawhlix sp.
Eue-pWw
sp.
Amber-
BolMgetM?
two sp.
Bp'
A species of the angulatum, zone.
Probably a new speoiea.
Probably a new species. The genus
is confined to the Lias.
One is probably a new sp.
Impreaaioh of base,
BRISTOL CHANNEL AREA AND THEIR VERTEBRATE FAUNAS
28 1
APPENDIX 11.
FOSSILS
OBTdINED FSOY THE SILICIFIED ROCK
MARKED AS LIASON THE OLD SERIES
ONE INCH
GEOLOGICAL
SURVEYMAP No. 19.
(Identifications by R. V. Melville, Esq.)
Localities.
Specimens were collected from :
1. South-east of Badgers Cross, a mile south-east of Gurney Slade. The rock
probably forms a disrupted cover in this area. When seen in an excavation blocks
of the rock were embedded in the subsoil, and were very numerous. No blocks of
limestone or Trim were present in the excavation, though the underlying solid rock
in the area is Carboniferous Limestone. The exterior of the blocks was stained
brown, presumably due to subsoil waters. Numerous blocks of the same rock occur
in hedgerows, farm gateways, etc. in the area and all have the stained exterior
except where they have been trimmed. No quarry of this stone exists in the Badgers
Cross area so presumably all the blocks have been obtained from the subsoil in the
locality. Specimens have been deposited a t the Geological Survey.
2. Highcroft Quarry, about one-third of a mile south-west of Gurney Slade.
Occasional blocks of rock are seen in the subsoil above the quarry face. Specimens
have been desposited a t the Geological Survey. The exterior of the blocks are
stained brown. The blocks are sparser in their distribution in the Highcroft Quarry
area than in the Badgers Cross area.
LAMELLIBRANCHIA
Trigonia bella Lycett
Myophorella etriata (J.Sowerby)
Myophorella (Vaugonia) angulata (J.de C . Sowerby)
Eopecten cf. velatua (Goldfuss)
Plagiostoma cf. aemicircularis (Goldfuss)
Astarte cf. elegans (J.Sowerby)
Gervillella of. muta (J. Sowerby)
Lopha costah (J.de C . Sowerby)
Camptumctu r i g i d w (J.Sowerby)
Liwmtula gibbosa (J. Sowerby)
i
Forms
characteristic
of the Inferior
Oolite.
GASTROPODA
Trochotoma cf. gradua (Deslongchamps)
BRACHIOPODA
large sulcate Terebratulas
Liassic Terebratulas are non-sulcate.
ECHJXODERBUTA-CRINOIDEA
Isocrinus sp.
EXPLANATION
OF THE
PLATES.
PLATE1 (5).
Photograph of the first cavern of the Slickstones Quarry underground watercourse, now cut. across
in transverse section near position D in text-figure 5 (11). Viewed from the sout,h.
PLATE 1 (b).
Outline tracing of the photograph in Plate I a, to show the boundary between the Carboniferous
Limeetone and the Triasaic sediments of the cavern.
1. Sediments of the oavern.
2. Part of the sediments of the narrow channel (see fig. 6) leading to the second cavern.
3. Carboniferous Limestone forming cavern W 8 b and roof.
4. Carboniferous Limestone of the main (northern) face.
6. Cerboniferoua Limestme of the northern part of the emtern fece (see fig. 6).
282
TBE
msozorc mssrrms OF
THE BRISTOL UHANNEL AREA
PLATE2 (a).
Photograph of tho N.W.face of t,he main Mesozoic deposit, Emborough Quarry.
PLXm 2 (b).
Traced Diegram, taken from the photograph of the N.W. face of the main Mesozoic deposit,
Emborough Quarry.
1. Limestone it& sitit..
2. Quamyrubbish.
3. E x p m of homogenous clay.
4. Conglomerate.
PLBTE 3 (a).
Phobograph of the S.W. face of the main Mesozoic deposit, Emborough Quarry.
PLATE
3 (b).
Traced diagram, t.aken from the photograph of the S.W. face of the main i\lesozoic
deposit, Emborough Quarry.
1. Pebbly stream deposit, completely suirroundcd by limestone.
2. Homogenous clay, about ten feet exposed.
3. Conglomerate of the N.W. face.
4. Limestone in situ.
5. Limeetone of the main faw.
6. Smell water-woG swirl-holes in the limestone filled with well-laminated clay.
7. Water-worn bedding plane.
PLATE
GEOLOGYOF
TEE
4.
ABEA N E ~ REMBOROU~H
QUARRIES.
Palaeozoic boundaries &r F. B. A. Welch (1929).
Mesozoic boundaries by the author. These differ from t,hose on the Geological Survey 1-inch
map, sheet 19 (Old Series), mainly as follows :1. The age of the outlier of Mesozoic rocks in the S.E. corner of the map has been shown
to be Inferior Oolite, not Lias (see Appendix 11).
2. An outlier of cherty Lias is ndicated on the Survey map, occupying much of the
ground between the Emborough Thrust and the Slab House Fault. On my map
its boundary would run approximately from the G. in Emborough Thrust to the
F in Fault, then dong the fault as far as the 700’ contour, there curving bmk to the
Emborough Thrust so as to include the patch of Tea Green Marls.
Fragments of siliceous rocks were found in ploughed fields adjacent to the T h t ,
aad larger pieces in walls forming field boundaries there and as far east as the Slab
House Fault. Them rocks divide naturelly into two groups. One consists of quartzite
and sandstones, similar in lithology to those seen in the quarry in the Carboniferous
pita W. of the ochre pockets. There seems no reason to suppose that these fra
are not Carboniferous, farmers having used material from the quarry for
repairing walls. The second type of siliceous rock commonly has a dark blue compact
‘heart ’,and a buff or cream-coloured‘ skin ’ of varying thichess. Some large boulders
of this type, obtained from the subsoil of one of the ochre pockets, contained Carboniferous Limeatone fossils, which showed no sign of being derived fragments. Possibly
the ailidcation, which has been shown to have affected rocks varying in age from
Triaaaic to Inferior Oolite in this part of the Mendips, has affected the Carboniferous
Limestone at this locality.
The only Meaozoic rocks which can be identified in this area are the ochre pockets
and the two patches labelled Tea Green Marls. The age of the ochre pockets is uncertain, as no foeails could be found in them, but, Morgan & Reynolds (1899)recorded
ochre pockets within and below the Triassic conglomermtes and sandstones formerly
exposed near the Emborough Quamea offices. The patches labelled Tea Green Mark
are bluish green clays which weather yellow, and form a very sticky clay soil. At the
southern tip of the northern outcrop these clays may be seen resting direct!y on the
Cerboniferoue Limestone. A similar clay, resting on red marls and followed by the
b a d Rhaetic bone bed, is recorded by Morgan & Reynolds from the pits formerly
worked near the Embomugh Quarries offices.
built::
ROBINSON
JOURN. LINN. SOC., ZOOL., VOL. XLIII. PLS. I and 2.
PLATE
1 (cr).-The Slickstone Quarry fissure : the first cavern, a t position D in fig. 5 ,
now cut, across by quarrying and viewed from the South.
PLATE
2 (a).-The
Embomugh Quarry fissure : the north-west side.
a .
ROBINSON.
JOURN. LINN. SOC. ZOOL., VOL. XLIII. PL. 3.
K - ZONE
Z -ZONE
C-ZONE
S-ZONE
D -ZONE
GRITS